Structure-activity relationships of small molecule autotaxin inhibitors with a discrete binding mode

Article abstract of DOI:10.1021/acs.jmedchem.6b01597

Autotaxin (ATX) is a secreted enzyme responsible for the hydrolysis of lysophosphatidylcholine (LPC) to the bioactive lysophosphatidic acid (LPA) and choline. The ATX-LPA signaling pathway is implicated in cell survival, migration, and proliferation; thus, the inhibition of ATX is a recognized therapeutic target for a number of diseases including fibrotic diseases, cancer, and inflammation, among others. Many of the developed synthetic inhibitors for ATX have resembled the lipid chemotype of the native ligand; however, a small number of inhibitors have been described that deviate from this common scaffold. Herein, we report the structure-activity relationships (SAR) of a previously reported small molecule ATX inhibitor. We show through enzyme kinetics studies that analogues of this chemotype are noncompetitive inhibitors, and by using a crystal structure with ATX we confirm the discrete binding mode.

Full text of DOI:10.1021/acs.jmedchem.6b01597

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES
Article
Structure-activity Relationships of Small Molecule
Autotaxin Inhibitors with a Discrete Binding Mode
Lisa M Miller, Willem-Jan Keune, Diana Castagna, Louise C. Young, Emma L Duffy, Frances
Potjewyd, Fernando Salgado-Polo, Paloma Engel García, Dima Semaan, John M. Pritchard,
Anastassis Perrakis, Simon J. F. Macdonald, Craig Jamieson, and Allan J. B. Watson
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01597 • Publication Date (Web): 16 Dec 2016
Downloaded from http://pubs.acs.org on December 18, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Structureꢀactivity Relationships of Small Molecule Autotaxin Inhibitors with a Discrete Binding
Mode
a
b
a
c
a
Lisa M. Miller, Willem-Jan Keune, Diana Castagna, Louise C. Young, Emma L. Duffy,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
b
a
b
Frances Potjewyd, Fernando Salgado-Polo, Paloma Engel García, Dima Semaan, John M.
d
b
d
,a
Pritchard, Anastassis Perrakis, Simon J. F. Macdonald, Craig Jamieson,* and Allan J. B.
,
a
Watson*
a
WestCHEM, Department of Pure and Applied Chemistry, University of Strathclyde, Thomas
Graham Building, 295 Cathedral Street, Glasgow, UK.
b
c
d
Division of Biochemistry, Netherlands Cancer Institute/Antoni van Leeuwenhoek, Plesmanlaan
1
21, 1066 CX, Amsterdam, Netherlands
Strathclyde Institute of Pharmacy and Biomolecular Science, University of Strathclyde, John
Arbuthnott Building (Hamnet Wing), 161 Cathedral Street, Glasgow, UK
Fibrosis Discovery Performance Unit, GlaxoSmithKline, Medicines Research Centre, Gunnels
Wood Road, Stevenage, UK
ABSTRACT
Autotaxin (ATX) is a secreted enzyme responsible for the hydrolysis of lysophosphatidylcholine
(
LPC) to the bioactive lysophosphatidic acid (LPA) and choline. The ATXꢀLPA signalling
pathway is implicated in cell survival, migration, and proliferation; thus, the inhibition of ATX is
a recognized therapeutic target for a number of diseases including fibrotic diseases, cancer, and
inflammation, amongst others. Many of the developed synthetic inhibitors for ATX have
resembled the lipid chemotype of the native ligand; however, a small number of inhibitors have
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
been described that deviate from this common scaffold. Herein, we report the structureꢀactivity
relationships (SAR) of a previously reported small molecule ATX inhibitor. We show through
enzyme kinetics studies that analogues of this chemotype are noncompetitive inhibitors, and
using a crystal structure with ATX we confirm the discrete binding mode.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
INTRODUCTION
Autotaxin (ATX) is a secreted glycoprotein phosphodiesterase belonging to the ectonucleotide
1
pyrophosphatase and phosphodiesterase family (ENPP). ATX mediates the hydrolysis of
lysophosphatidylcholine (LPC) 1 to the ubiquitous bioactive lipid mediator lysophosphatidic
2
acid (LPA) 2 and choline 3 (Scheme 1). 2 acts on the LPA receptors; a series of Gꢀprotein
coupled receptors (GPCRs), of which six have been identified to date, termed LPA , with an
1
ꢀ6
additional three pending the publication of evidence to verify their categorization (GPR87,
3
P2Y10, and GPR35). Upon binding, a series of intracellular signaling pathways are invoked
which could result in a range of cellular functions including differentiation, migration,
4
proliferation, and survival. Deregulation of the ATXꢀLPA signaling pathway is implicated in a
variety of disease states and could lead to the development of several diseases including
vascularization diseases, autoimmune diseases, cancer, fibrotic diseases, inflammation, and
5ꢀ10
neurodegeneration, amongst others.
Accordingly, the use of ATX inhibitors has been and
continues to be an attractive strategy for therapeutic intervention (for a recent review see
1
1
Castagna et al. ).
Scheme 1. ATXꢀMediated Hydrolysis of LPC Releasing LPA: Disease States via LPA Action on
the GPCRs LPA . R group = fatty acid chain with varying degrees of unsaturation.
1
ꢀ6
ACS Paragon Plus Environment

Page 3 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
With respect to the endogenous ligand (1), it is perhaps unsurprising that many of the synthetic
ligands developed for ATX have resembled the lipid chemotype: an acidic ‘head’ group linked
12
through a variety of ‘cores’ to a lipophilic ‘tail’ (Figure 1: 4, 5, 6, and 7). However, a
comparatively smaller number of inhibitors have been described that deviate from this common
1
3
scaffold, for example PATꢀ347 (8), reported by PharmAkea Therapeutics. Lipidꢀbased
inhibitors have been previously reported to bind within the catalytic site of ATX and thus
1
7
compete for the same region that the natural ligand is known to bind; however, a recent report
from PharmAkea found the small molecule inhibitor 8 to be a potent noncompetitive inhibitor of
1
3
ATX, with a distinct mode of binding remote to the catalytic site. This compound was noted to
have a scaffold similar to our compound of interest, the indoleꢀderived inhibitor 9 originally
18
reported by Amira Pharmaceuticals. The patent in which 8 and related analogues are claimed
1
3
report the IC values in ≤0.5 µM, >0.5 but ≤3 µM, or >3 µM categories; similarly, the patents
50
in which analogues of 9 are reported also categorize the activities (IC <300 nM, 0.3ꢀ1 ꢁM, or
5
0
18
>
1 ꢁM). From consideration of this data very little detailed SAR of these nonꢀlipidꢀlike ATX
inhibitors can be extracted. Based on this, we generated a focused set of 50 analogues to explore
the SAR of 9 in greater detail and to determine the features which contribute to its potency.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
4
15
16
17
13
Figure 1. LipidꢀBased Inhibitors and Small Molecule Inhibitors of ATX: 4, 5, 6, 7, 8, 9
18 a
b
(
previously reported activity from patent literature). FSꢀ3 assay; LPC choline release assay.
Herein, we report the design, synthesis, and biological evaluation of our compound collection
and describe the main determinants of potency regarding these indoleꢀbased compounds. We
show that this chemotype has only limited activity in a bisꢀ(pꢀnitrophenol) phosphate (bisꢀpNPP)
based assay; however, using an LPC choline release assay a number of potent inhibitors were
identified. Analysis of this data, combined with further kinetic studies into the mode of ATX
inhibition, has furnished a new insight into the binding site of this chemotype. Furthermore, we
report the crystal structure of ATX in complex with an analogue of 9, confirming that this
chemotype binds in the same discrete binding site identified by PharmAkea in the cocrystal of 8
1
3
with ATX.
ATX Inhibition Assays. There are a number of in vitro assays that are currently used to evaluate
ATX activity; these can be divided into two categories based on the type of substrate the assay
utilizes (Table 1). The first category employs the natural ligand 1 and the second uses unnatural
ATX substrates. As depicted in Scheme 1, ATX hydrolyzes 1 into 2 and 3, thus the enzyme
ACS Paragon Plus Environment

Page 5 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
activity can be determined by quantitative measurement of these products using liquid
1
9
chromatographyꢀtandem mass spectrometry (LCMS), radiometry, colorimetry, or fluorimetry.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. Summary of ATX inhibition assays.
Category
Substrate
Analysis
1
9a
19b
19c
Natural Substrate
LPC
LCMS,
radiometry,
colorimetry,
1
9d
fluorimetry
2
Unnatural Substrate
pNPꢀTMP
Absorbance
Absorbance
FRET
17
bisꢀpNPP
2
0
CPF4
2
1
FSꢀ3
Fluorescence
Fluorescence
22
TGꢀmTMP
Of assays based on 1, a commonly used method measures the production of 3 using an enzymeꢀ
1
9c
coupled colorimetric assay
– choline oxidase converts 3 into betaine 10 and hydrogen
peroxide, and then horseradish peroxidase (HRP) utilizes the hydrogen peroxide to oxidize a
coloring substrate into its chromophoric state (Scheme 2). There is a risk of false positives
associated with this assay as the molecules under analysis may inhibit the enzymes used in the
23
coloring reactions; however, control reactions can be employed to reveal any interference.
Scheme 2. ATXꢀmediated choline release colorimetric assay for ATX Inhibition. R group = fatty
acid chain with varying degrees of unsaturation.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Accordingly, assays that employ synthetic substrates can appear to be advantageous over the
LPC choline release assay as they do not require the use of these additional enzymes. There are a
number of unnatural substrates widely used, including the nucleotide thymidine
2
5’monophosphate pꢀnitrophenyl ester (pNPꢀTMP), the Förster resonance energy transfer
2
0
21
(
FRET) ligand CPF4, the fluorogenic substrate 3 (FSꢀ3), and the fluorescence probe
2
2
TokyoGreen mꢀthymidine monophosphate (TGꢀmTMP). A particularly widely used method
employs the unnatural substrate bis(pꢀnitrophenyl) phosphate (bisꢀpNPP) 11. ATX catalyses the
hydrolysis of 11 to release 4ꢀnitrophenol 12 which, with an absorbance of 405 nm, is readily
detected using colorimetry (Scheme 3). The popularity of this assay can be attributed to the low
1
7
cost of this substrate and the direct readout the assay provides.
Scheme 3. ATXꢀMediated Hydrolysis of BisꢀpNPP 11 Releasing 4ꢀNitrophenol 12.
ACS Paragon Plus Environment

Page 7 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The recent report from PharmAkea Therapeutics described four structurally distinct ATX
inhibitors that showed good inhibition using the LPC choline release assay but significantly
1
3
lower potency when tested using the bisꢀpNPP or FSꢀ3 substrates. The differences in datasets
of the three assays used were found to be a result of alternative binding modes within ATX,
which was confirmed by crystallographic data identifying proximal binding sites. As stated
above, these compounds were noted to have a scaffold similar to our compound of interest, the
indoleꢀderived inhibitor 9 originally reported by Amira Pharmaceuticals. In the choline release
assay employed by Amira Pharmaceuticals, 9 is claimed to show an IC of <300 nM, however,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5
0
1
8
no further detail was given. In addition, 9 was reported to have activity in a mouse air pouch
assay, collagenꢀinduced arthritis, a rat model of neuropathic pain, and in lung melanoma
metastasis, although the efficacy of 9 in these assays was not disclosed. Herein, we report that 9
and related analogues were found to display only limited activity in the bisꢀpNPP assay;
however, using the LPC choline release assay a number of efficient inhibitors were identified,
with IC values ranging from 4 nM to >30 ꢁM for the complete series. The disconnect observed
5
0
between the two assay datasets for this chemotype were hypothesized to be due to an alternative
binding mode, remote to the catalytic site. We report enzyme kinetic studies and a coꢀcrystal
structure of ATX with an analogue of 9 to support this alternative binding site. Furthermore, we
also report the exploration of the SAR associated with this chemotype.
Molecule Design. Using the limited data available, the key structural motifs of 9 were suspected
to be (i) the chloro substitution at the 6ꢀposition of the indole, (ii) the substitution of the indole
nitrogen, and (iii) the thiopyridine carboxylic acid. From this analysis we were able to infer the
initial SAR shown in Figure 2. In the current study, we aimed to probe this emerging SAR
further, to test our observations, and to more completely establish the contribution of each of
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
these principal structural motifs towards potency through the synthesis and biological evaluation
of a series of analogues.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
8
Figure 2. SAR of indole 9 extracted from data reported by Amira Pharmaceuticals.
The first region of interest was the 6ꢀCl substitution of the indole ring. Analysis of the patent
data had suggested this was an important region of the molecule, with changes to substituent or
position on the ring having a direct effect on the activity observed. To explore this relationship in
greater depth, a series of compounds were designed with alternative chloro regiochemistry, as
well as removal or exchange of the chloro for alternative functional groups. Three alternatives
were chosen to determine the optimal size and nature of the substituent: fluoro, methoxy, and
methyl. The second region to be modified was at the Nꢀ1 position of the indole ring. The Nꢀ
methyl pyrazole was replaced with a phenyl ring, cyclopentyl, and methyl to explore the
tolerance of this region for a less polar aromatic ring as well as aliphatic groups. Finally, the
thiopyridine carboxylic acid was modified, with changes made to the linker and the nature of the
aromatic ring. The previously reported data in the patent literature contained little SAR around
the linker area of the compound and therefore we were interested in converting the thioether to
an ether or methylene linker, as well as exploring the difference between pyridyl and phenyl
analogues.
Chemistry. The known ATX inhibitor indole 9 was prepared as shown in Scheme 4. Starting
2
4
from ketone 13, the indole was synthesized using a Fischer synthesis as previously reported,
ACS Paragon Plus Environment

Page 9 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
18
either according to or modified from literature procedures. This reaction produced a mixture of
the 4ꢀCl and 6ꢀCl substituted indoles, which were successfully separated on silica to isolate the
desired 6ꢀCl regioisomer 16 as the major product. The 4ꢀCl analogue 14 was also carried through
the synthesis as part of the SAR study around the indole substitution. Introduction of the 1ꢀ
methylꢀpyrazole substitution to the indole nitrogen proved to be challenging, requiring harsh
conditions and long reaction times during which partial hydrolysis of the ester occurred. As a
result, the substitution reaction was telescoped into the final step to fully hydrolyze the ester and
produce the desired acid 9.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Scheme 4. Synthesis of the Amira Compound 9.
a
Reagents and conditions: (a) (i) tBuOH, 3ꢀchlorophenylhydrazine hydrochloride, reflux; (ii)
HCl, AcOH; (b) 2 M aq. NaOH, THF, RT; (c) 4ꢀiodoꢀ1ꢀmethylpyrazole, K CO , CuO, pyridine,
2
3
170 °C.
Initial syntheses followed the reported methods from the Amira Pharmaceuticals patents to
produce the desired indole 16 in good yield (71%, see experimental section) (Scheme 4).
However, in the exploration of this key step it was discovered that a number of the Fischer
indole reactions proceeded by refluxing the ketone with the respective hydrazine hydrochloride
in EtOH without an additional acid catalyst (see experimental section). The synthesis of
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
analogues in which the chloro was exchanged for an alternative functionality (F, OMe, and Me)
was achieved using the same synthetic strategy employed in the synthesis of compound 9. Using
this synthetic route, the analogues described in Figure 3 were accessed. Unfortunately, synthesis
of the 7ꢀCl analogue with the 1ꢀmethylꢀpyrazole Nꢀsubstituent was not achieved, as efforts to
introduce the 1ꢀmethylꢀpyrazole were unsuccessful.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Substitution on the indole ring.
The next series was designed to investigate the SAR associated with the 1ꢀmethylꢀpyrazole
substituent on the indole nitrogen. In place of the 1ꢀmethylꢀpyrazole three alternative substituents
were used; methyl, cyclopentyl, and phenyl. The aliphatic substituents were introduced by an
alkylation reaction using the corresponding alkyl halide and NaH (Scheme 5).
a
Scheme 5. Synthesis of Compound 37.
ACS Paragon Plus Environment

Page 11 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
Reagents and conditions: (a) (i) tBuOH, phenylhydrazine hydrochloride, reflux; (ii) HCl,
AcOH; (b) NaH, MeI, THF, reflux; (c) 2 M aq. NaOH, THF, RT.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Following the alkylation step, hydrolysis of the corresponding ester revealed the desired
carboxylic acid. The phenyl substituent was introduced using an alternative Ullmann coupling
and the 1ꢀmethylꢀpyrazole was introduced using the conditions described previously. Using these
approaches, the analogues described in Figure 4 were successfully synthesized.
Figure 4. Substitution on the indole nitrogen.
The benzoic acid compounds were synthesized in a similar way to the pyridyl series with an
additional step to introduce the ester functionality using a methoxycarbonylation reaction
(Scheme 6). From this intermediate the desired alkyl or aryl Nꢀsubstituent was introduced,
followed by hydrolysis to gain the desired acid 47 in the final step of the synthesis. The ester 46,
and its deschloro counterpart 92 (not shown, see experimental section), then allowed access to
the analogues described in Figure 5.
a
Scheme 6. Synthesis of Compounds 45ꢀ48.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
Reagents and conditions: (a) chloroacetone, DIPEA, MeCN, reflux; (b) 3ꢀ
chlorophenylhydrazine hydrochloride, EtOH, reflux; (c) Herrmann’s catalyst, Mo(CO)6,
[tBu PH]BF , DBU, MeOH/MeCN (1:4), 70 °C; (d) NaH, MeI, THF, reflux; (e) 2 M aq. NaOH,
3
4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
THF, RT.
Figure 5. Substitution on the indole nitrogen, benzoic acid analogues.
The oxygenꢀlinked benzoic acid subset (Figure 6) were synthesized using the Fischer indole
reaction that was employed with previous analogues (Scheme 7). Efforts to synthesize the
corresponding ether linked pyridyl compounds were unsuccessful (see Electronic Supporting
Information (ESI)).
a
Scheme 7. Synthesis of Compounds 59ꢀ62.
a
Reagents and conditions: (a) chloroacetone, K CO , acetone, reflux; (b) phenylhydrazine
2
3
hydrochloride, EtOH, reflux; (c) 2 M aq. NaOH, THF, RT; (d) 4ꢀbromoꢀ1ꢀmethylpyrazole,
K CO , CuO, pyridine, 170 °C.
2
3
ACS Paragon Plus Environment

Page 13 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Ether linked analogues.
In order to access the methylene linked compounds an alternative synthetic strategy was required
as efforts to employ a Fischer indole synthesis were unsuccessful. It was discovered that by
reordering the synthetic steps to install the 1ꢀmethylꢀpyrazole at an earlier stage, the Nꢀ
substitution proceeded in a much improved yield than the late stage introduction used in previous
syntheses (Scheme 8). A twoꢀstep alkylation of 66 using 6ꢀbromopicolinaldehyde afforded the
desired methylene linker in good yield. From the bromo intermediate 97 (not shown) the ester
functionality was introduced via a methoxycarbonylation reaction to give 67, followed by
hydrolysis to reveal the desired acid 68. Thus, this strategy was used to access the remaining
compounds in this series (Figure 7).
a
Scheme 8. Synthesis of Compounds 66ꢀ68.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
Reagents and conditions: (a) 4ꢀbromoꢀ1ꢀmethylpyrazole, K CO , CuO, pyridine, 170 °C; (b) (i)
2
3
6
ꢀbromopicolinaldehyde, CH Cl (ii) Et SiH, TFA, CH Cl ; (c) Herrmann’s catalyst, Mo(CO) ,
2 2 3 2 2 6
[tBu PH]BF , DBU, MeOH/MeCN (1:4), 70 °C; (d) 2 M aq. NaOH, THF, RT.
3
4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 7. Methylene linked analogues.
To determine the importance of the indole core, this motif was removed and compound 75 was
prepared featuring a phenyl at its center but retaining the 1ꢀmethylꢀpyrazole motif. This
compound was obtained from the commercial carboxylic acid 72 in four steps (Scheme 9). This
required an alternative strategy, using a SuzukiꢀMiyaura reaction to introduce the 1ꢀmethylꢀ
pyrazole via the coupling of 74 with 1ꢀmethylpyrazoleꢀ4ꢀboronic acid pinacol ester.
a
Scheme 9. Synthesis of Analogue 75.
a
Reagents and conditions: (a) conc. H SO , MeOH, reflux; (b) 4ꢀbromothiophenol, DMF, 45 °C;
2
4
(
c) 1ꢀmethylpyrazoleꢀ4ꢀboronic acid pinacol ester, K PO , Pd(OAc) , SPhos, THF/H O (7:1), 50
3 4 2 2
°C; (d) 2 M aq. NaOH, THF, RT.
ACS Paragon Plus Environment

Page 15 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The final deletion analogue was designed to remove one of the aromatic rings from the
compound and replace it with an aliphatic carboxylic acid (Scheme 10). Starting from 4ꢀ
bromobutyric acid 76 the thiol group was introduced by an Sꢀalkylation of thiourea, which was
subsequently hydrolyzed in situ to form the primary thiol 77. This was then followed by a second
alkylation using chloroacetone to access the Fischer indole precursor 78. With 79 in hand, the 1ꢀ
methylꢀpyrazole was introduced at the indole nitrogen (Scheme 10). This route was also
employed to access the 6ꢀCl analogue 81 (Figure 8).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Scheme 10. Synthesis of Aliphatic Carboxylic Acid Analogue 80.
a
Reagents and conditions: (a) thiourea, EtOH, reflux; (b) chloroacetone, DIPEA, THF, RT; (c)
phenylhydrazine hydrochloride, EtOH, reflux; (d) 4ꢀbromoꢀ1ꢀmethylpyrazole, K CO , CuO,
2
3
pyridine, 170 °C.
Figure 8. Aliphatic acid analogues.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Results and Discussion. To evaluate our compound library, the first screen was performed using
the bisꢀpNPP assay. This was selected based on its broad application throughout the ATX
1
1
literature. An initial screen was carried out using 30 ꢁM of inhibitor allowing a rapid
evaluation of the compounds. The criteria of >60% inhibition was set for a compound to be
termed active and therefore progressed to further testing. Based on this initial survey of the
activity of Amira compound 9 along with the 49 analogues in the bisꢀpNPP assay, we identified
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
10 potential inhibitors, including the patent compound 9 (see ESI). These 10 compounds were
tested again in the bisꢀpNPP assay using a concentration range of 30 nM – 30 ꢁM inhibitor, to
determine the IC for each of these compounds. Using the bisꢀpNPP assay, the Amira ATX
5
0
inhibitor 9 was found to have an IC of 22 nM, which was consistent with the value of <300 nM
5
0
1
8
claimed in the patent (LPC choline release assay); however, the IC values observed in this set
5
0
ranged from 22 nM to >30 ꢁM (Table 2). These results demonstrated that our first screen at an
inhibitor concentration of 30 ꢁM had not been a reliable indicator of activity. Importantly, on
further inspection of the dose response curves (Figure 9), it was noted that when using the bisꢀ
pNPP assay, compound 9 showed only 66% maximal inhibition of ATX at 30 µM (34%
hydrolysis of 11), indicating that this analogue was not fully inhibiting the ATX hydrolysis of
11, potentially implying a discrete binding mode compared to that of the artificial substrate.
ACS Paragon Plus Environment

Page 17 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 9. Testing of 9 in the bisꢀpNPP assay from 30 nM – 30 ꢁM of inhibitor. Dashed line
indicates 60% inhibition criteria of initial screen. Control reactions confirmed there was no
hydrolysis of 11 in the absence of ATX (data not shown).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A further 20 analogues were tested in the bisꢀpNPP assay confirming that partial inhibition was a
consistent feature associated with this series (data not shown, see ESI). Although the bisꢀpNPP
assay has many advantages, there are risks associated with the use of unnatural substrates if the
binding interactions with the target enzyme are dissimilar to those of the natural substrate. There
are a number of reported examples where misleading results were observed when using ATX
17,25,26
assays that employ the unnatural ligands FSꢀ3, CPF4, and pNPꢀTMP.
The authors
concluded that the discrepancies in assay datasets were the result of the inhibitor compounds and
assay substrates binding to different regions of ATX. It has been well documented that the active
site of ATX features three key regions with which small molecule inhibitors can interact: the
2
7ꢀ29
catalytic site, the hydrophobic pocket, and the tunnel.
Comparing the likely binding of the
unnatural substrate 11 with that of ATX in complex with 2, it is noted that while both bind at the
catalytic site, the acyl chain of 2 shows additional extensive interactions with the hydrophobic
pocket (Figure 10). This suggests that it is possible for small molecules to bind outside the
catalytic site without inhibiting, or only partially inhibiting, the hydrolysis of 11, potentially
resulting in false negative results in the bisꢀpNPP ATX assay.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 10. Three binding sites of ATX: Catalytic Site, Hydrophobic Pocket, and Tunnel.
Complex of LPA 14:0 (red) in mouse ATX (PDB 3NKN) with the docked position of bisꢀpNPP
3
0
31
1
1 (green) overlaid. Docked using GOLD and viewed using Discovery Studio Visualizer.
Indeed, the noncompetitive inhibitor 8 from PharmAkea, which shares a number of structural
13
similarities to 9, has been reported to bind an allosteric site within the tunnel. The authors
observed that 8 inhibited the hydrolysis of 1 but not of 11, showing that an inhibitor binding
within the tunnel may not be detected using the bisꢀpNPP assay. Moreover, selected steroid
analogues have recently been shown to bind the same site, with the activity of ATX in
2
9
hydrolyzing LPC allosterically inhibited but not its activity on pNPꢀTMP. These reports led us
to employ the natural substrate 1 in the LPC choline release assay to further evaluate our
inhibitors, gain a more complete understanding of the underlying SAR, and offer insight into
potential binding modes. Prior to testing the inhibition of ATX, a control reaction was run using
three exemplar compounds to ensure that there was no incompatibility with the two coupling
ACS Paragon Plus Environment

Page 19 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
enzymes used in the assay: choline oxidase and HRP. Using choline chloride as the substrate, no
inhibition was observed for the latter two reactions of the LPC assay, as shown in Scheme 2 (see
ESI). Testing proceeded with an initial screen of all 50 compounds from 400 nM – 100 ꢁM, to
identify active inhibitors. From this dataset, compounds that exhibited an IC <1 ꢁM were tested
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
50
further across a concentration gradient of 0.3 nM – 10 ꢁM, and the IC values were determined
50
(compounds with IC50 >30 ꢁM not shown, see ESI). This assay provided us with robust
inhibition data, with the more potent compounds showing full inhibition of ATX. Of the 50
analogues described in the current study, only four had been previously reported by Amira
1
8
Pharmaceuticals (9, 47, 51, 52). The IC values determined in our LPC choline release assay
5
0
were consistent with the values claimed in the patents (Table 2). It is also important to note that
substrate concentration can affect the IC value observed in an assay; for competitive assays
5
0
such as bisꢀpNPP and the LPC choline release assay, the substrate concentration should be less
3
2
than, or equal to, the MichaelisꢀMenten constant (K ) for the given substrate. The biological
M
evaluations detailed in this current study were carried out using substrate concentrations equal to
the K values of 1 (~150 µM) and 11 (~1 mM) for ATX previously reported by van Meeteren et
M
33
al.
Table 2. BisꢀpNPP and LPC assay results.
BisꢀpNPP
LPC
Compound
9
IC (nM)
IC (nM)
5
0
50
a
22 ±6.9
4 ±0.5
17
23
24
28
604 ±113
25 ±13
b
ꢀ
>20 µM
b
ꢀ
>8000
5700 ±2500
1700 ±400
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
9
0
ꢀ
ꢀ
258 ±127
b
3
>3500
31
816 ±564
ꢀ
349 ±125
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
b
3
3
3
4
2
3
5
0
>10 µM
85 ±22
75 ±37
25 ±12
124 ±37
31 ±11
b
>4000
b
41
2000 ±785
ꢀ
>2800
c
4
2
7
ꢀ
d
4
100 ±36
290 ±35
1200 ±800
747 ±500
49
b
b
50
51
52
56
64
68
69
>4000
>20 µM
a
5 ±2.5
ꢀ
81 ±28
a
11 ±7
b
>30 µM
204 ±88
89 ±43
33 ±16
>5000
726 ±155
339 ±190
660 ±224
b
80
ꢀ
>10 µM
a
All tested from 0.3 nM – 10 ꢁM, 10 point curves, n = 2. Consistent with the value of < 300 nM
b
c
claimed in the patent, LPC assay. Incomplete curve observed. Unable to determine IC due to
5
0
d
solubility issues in the assay medium. Consistent with the value of >1 uM claimed in the patent,
LPC assay.
ACS Paragon Plus Environment

Page 21 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Structureꢀactivity Relationship. Using the results from the LPC assay we were able to define
the SAR around this chemotype and provide a better understanding of the features which
contribute to its potency. During our examination of the patent data, it became apparent that the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
6
ꢀCl of compound 9 was of high importance to the potency. The deletion analogue 75, in which
the indole was removed with a phenyl ring used as the core, had resulted in a complete loss of
activity (data not shown, see ESI), thus confirming the importance of the indole core. To
investigate the SAR of the 6ꢀCl substitution on the indole, a series of analogues were made
varying the position of the chloro around the benzenoid ring, as well as removing the substituent
(Table 3). We also aimed to explore the significance of the chloro by exchanging this for
alternative functionalities: F, OMe, and Me. The effect of this chloro on potency was apparent
throughout our results, with its presence increasing the potency in every example. It was also
identified that the 6ꢀposition of the indole ring was optimal for the chloro, with the other
regioisomers resulting in lowered or a complete loss of activity (Table 3). This can be noted in
the comparison of compounds 9, 18, and 29, with the movement of the chloro substituent around
the indole ring.
S
N
OH
R
Me
O
N
N
N
Me
Table 3. SAR of indole ring.
Compound
R
IC (LPC)
50
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
28
H
1700 ±400 nM
>30 ꢁM
1
2
8
9
4ꢀCl
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5ꢀCl
258 ±127 nM
4 ±0.5 nM
9
6ꢀCl
a
30
4ꢀF
>3 µM
31
6ꢀF
349 ±125 nM
a
32
4ꢀMe
6ꢀMe
4ꢀOMe
6ꢀOMe
>10 µM
33
25 ±12 nM
>30 ꢁM
3
3
4
5
124 ±37 nM
a
Incomplete curve observed.
In order to rationalize the effect of the 6ꢀCl, docking studies were performed using the apo form
of the ATX enzyme. From this work it was proposed that the 6ꢀCl projects into a nearby
hydrophobic cleft (Leu160, Ala164, Trp207, Trp161), making a series of hydrophobic
interactions (Figure 11B). The 6ꢀCl substituent is a vital motif for potency throughout this
chemotype, therefore we hypothesize that the hydrophobic interactions of this substituent with
the hydrophobic cleft are key contributors to the potency. It was possible to exchange the 6ꢀCl
ACS Paragon Plus Environment

Page 23 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
for alternative substituents (F, 31; Me, 33; OMe, 35) and retain the activity, although the chloro
was optimal. Based on our in silico model, a conclusion was drawn that the 6ꢀCl substituent
provides the optimal size to fit within the hydrophobic pocket with a larger (Me, 33) or smaller
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(
F, 31) substituent resulting in a loss of potency.
A
B
Phe157
C
Figure 11. In silico studies created using coꢀcrystal of ATX with 2 and 8 PDB 4ZG7. A: Model
of 51 (yellow) overlayed with cocrystal of 8 (colored by atom type) and ATX, with 2 residing in
the catalytic site and hydrophobic pocket. B: Model of ATX with 9 highlighting possible
interactions. C: Shift observed when 4ꢀCl substituted 18 is docked (yellow) vs. 6ꢀCl substituted 9
30
31
(colored by atom type). Docked using GOLD and viewed using Discovery Studio Visualizer
(see ESI).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Moving these alternative substituents from the 6ꢀ to the 4ꢀposition of the indole resulted in loss
of activity, providing further support to the importance of the 6ꢀposition. It was noted that with
no substituent on the indole some weak activity was observed with an IC50 of 1700 nM for
compound 28; however, introduction of a substituent to the 4ꢀposition resulted in loss of activity
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(18, 32, 34), with the exception of the 4ꢀF which is tolerated (30), showing that substitution at
this position prevents binding of these compounds. Using our in silico model, we suggest that a
clash between the 4ꢀposition substituents and a nearby residue (Phe157) alters the binding of
these analogues (Figure 11B). This shift in the binding position caused by the clash with the
protein could explain the loss in potency observed when a substituent is introduced to the 4ꢀ
position.
The next region to be explored was the substitution on the indole nitrogen. In order to establish
whether the aromatic substituent was required for activity, and if the size of the substituent was
important, three alternative substituents were used in place of the 1ꢀmethylꢀpyrazole: methyl,
cyclopentyl, and phenyl. Analogues with the unsubstituted indole were also synthesized to
determine whether any substituent at this region was required for activity. Based on the patent
data, the 1ꢀmethylꢀpyrazole is important for activity and removal of this substituent did result in
a reduction in potency; however, compound 17 with no decoration of the indole nitrogen, was a
potent compound with an IC of 25 nM. Further analogues exploring the substitution of the
5
0
indole nitrogen (methyl, cyclopentyl, phenyl) found that the 1ꢀmethylꢀpyrazole remained the
most active (Table 4). Unfortunately, we were unable to determine the IC of compound 42 due
5
0
to poor solubility in the assay medium observed during the screening of this analogue. Our
binding model had indicated possible πꢀπ interactions for the 1ꢀmethylꢀpyrazole of 9 with nearby
residues (Trp207: pyrazole face to indole edge and Phe221: face to face). By changing the 1ꢀ
ACS Paragon Plus Environment

Page 25 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
methylꢀpyrazole to an aliphatic substituent these interactions are lost, thus resulting in a loss of
activity. The methyl group of 40 was well tolerated, whereas the cyclopentyl substituent of 41
was found to cause a large drop in potency. These results suggest that the larger aliphatic
substituent is detrimental, either through unfavorable interactions or by clashing with the protein;
however, we did not observe either from consideration of our binding model.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 4. SAR of indole nitrogen substituent, with 6ꢀCl.
1
2
Compound
9
X
N
N
N
N
N
N
R
R
IC (LPC)
50
Cl
Cl
Cl
Cl
Cl
H
1ꢀmethylꢀpyrazole
4 ±0.5 nM
25 ±13 nM
31 ±11 nM
1
7
0
H
4
Me
b
41
cyclopentyl
>2800 nM
a
42
Ph
H
ꢀ
19
>30 ꢁM
1700 ±400
nM
28
N
H
1ꢀmethylꢀpyrazole
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
37
N
N
H
Me
>30 ꢁM
>30 ꢁM
38
39
52
49
H
H
cyclopentyl
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N
Ph
>30 ꢁM
CH
CH
Cl
Cl
1ꢀmethylꢀpyrazole
H
11 ±7 nM
747 ±500 nM
1
200 ±800
47
CH
Cl
Me
nM
b
5
0
1
CH
CH
Cl
Cl
cyclopentyl
>25 µM
5
Ph
81 ±28 nM
a
b
Unable to determine due to solubility issues in the assay medium. Incomplete curve observed.
The relationship between the 1ꢀmethylꢀpyrazole and the 6ꢀCl was highlighted by altering the
substitution of the indole nitrogen in a series of compounds that did not feature the 6ꢀCl
substitution. In this series of analogues (Table 4), only one compound exhibited weak activity,
which was the 1ꢀmethylꢀpyrazole 28, thus showing that either the 6ꢀCl or the 1ꢀmethylꢀpyrazole
must be present for activity; however, the 6ꢀCl substituent has the most profound effect on
activity.
This series of compounds was explored further via the synthesis of the analogous series with the
nitrogen removed from the pyridine ring, thus allowing for the SAR of the pyridyl acid and the
ACS Paragon Plus Environment

Page 27 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
phenyl to be determined. The benzoic acid analogues demonstrated the same SAR as the pyridyl
subꢀseries, with the 1ꢀmethylꢀpyrazole (52) giving the best inhibition (Table 4). Fortunately,
compound 51 showed good solubility in the assay medium, allowing for the IC50 to be
determined at 81 nM, unlike the pyridyl analogue 42 with which we had observed poor solubility
during screening in the LPC assay. This set of analogues additionally supports our observation
from the data in Table 4; that it is beneficial to have a substituent of aromatic nature on the
indole nitrogen, and although a cyclopentyl is unfavorable (50), smaller aliphatic groups can be
tolerated (47).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The acidic group of this chemotype was also found to be vital for ATX activity, as shown by the
two deletion analogues in which the thioether linked pyridine was removed (66) or exchanged
for an aliphatic acid (81) which resulted in loss of activity (Table 5). During our docking studies
of 9, it was noted that both the indole ring and the 1ꢀmethylꢀpyrazole substituent could
potentially contribute to binding through πꢀπ interactions; however, with the pyridyl acid
orientated towards the solvent it made no specific interactions with ATX, and only a potential Hꢀ
bonding interaction with a crystallographic water molecule was identified. The loss of activity
observed with 66 and 81 could indicate that the aromaticity is required for additional interactions
not identified in our docking study, or that the increased number of rotatable bonds was
unfavorable in the binding of this compound within ATX. By comparing the benzoic acid
compound 52 with the corresponding pyridyl analogue 9 we observed no significant change in
potency with removal of the pyridyl nitrogen; however, by comparing matched pair examples
with Nꢀsubstituents other than the 1ꢀmethylꢀpyrazole, it can be noted that the pyridine in general
improves potency (Table 4).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 5. SAR of the carboxylic acid region.
Compound
9
R
IC (LPC)
50
4 ±0.5 nM
11 ±7 nM
>30 ꢁM
5
2
1
8
66
H
>30 ꢁM
The thioether linker was then explored through replacement with an ether or a methylene.
Previous reports from the patent literature had disclosed no analogues with the methylene, and
only five with an ether linker, thus we sought to fill the SAR gap with a set of analogues using an
ether and methylene with both the pyridyl and benzoic acids. A large drop in potency was
observed as a result of changing the thioether (9) to a methylene spacer (68). We also noted this
trend in the benzoic acid analogues with the thioether (52) being superior to an ether (64) or
methylene (69) linker, though both alternatives were tolerated (Table 6). Our docking model
provided little insight into the effects of changing the linker (see ESI); thus we hypothesize that
the longer bond lengths of the thioether are beneficial to the binding of this chemotype.
ACS Paragon Plus Environment

Page 29 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Furthermore, this subset again demonstrates the importance of the 6ꢀCl for activity and the
superiority of the pyridyl analogues over the corresponding phenyl analogues.
Y
X
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
OH
Me
O
N
R
N
N
Me
Table 6. SAR of the linker.
Compound
R
H
Y
S
X
N
N
C
C
C
C
N
N
C
IC (LPC)
50
28
9
1700 ±400 nM
4 ±0.5 nM
>30 ꢁM
Cl
H
S
5
5
6
6
7
3
2
2
4
1
S
Cl
H
S
11 ±7 nM
>30 ꢁM
O
Cl
H
O
726 ±155 nM
>30 ꢁM
CH
CH
CH
2
2
2
68
Cl
H
339 ±190 nM
>30 ꢁM
7
0
ACS Paragon Plus Environment